Empirical studies of Gaussian process based Bayesian optimization using evolutionary computation for materials informatics
作者：
Highlights：
• Evolutionary computation is applied to Bayesian optimization (BO).
• A goal-directed acquisition function is proposed for materials informatics.
• Experiments for materials search are conducted on several initial size data for BO.
• Experiments on initial small size data show the usefulness of random search with BO.
• Regret function of BO and minimum search steps of the random search are analyzed.
摘要
•Evolutionary computation is applied to Bayesian optimization (BO).•A goal-directed acquisition function is proposed for materials informatics.•Experiments for materials search are conducted on several initial size data for BO.•Experiments on initial small size data show the usefulness of random search with BO.•Regret function of BO and minimum search steps of the random search are analyzed.
论文关键词：Bayesian optimization,Markov random search,Gaussian process,Ensemble,Evolutionary computation,Materials informatics
论文评审过程：Received 10 July 2017, Revised 9 October 2017, Accepted 10 November 2017, Available online 15 November 2017, Version of Record 5 December 2017.
论文官网地址：https://doi.org/10.1016/j.eswa.2017.11.026
